How to fool the world with bad science (Synopsis)

“You can’t cheat an honest man. Never give a sucker an even break or smarten up a chump.” -W.C. Fields

Well, it was bound to happen. A bad science story -- a story of illegitimate science with all the right buzzwords -- has gone viral. By now, you, too, have probable heard about the "impossible" space engine that's been validated by NASA and could take us to Mars: the EmDrive.

Image credit: SPR, Ltd. Image credit: SPR, Ltd.

But did it really work? Could it work? Or is this another hallmark example of the worst in not only science reporting, but in bad science itself?

Image credit: “Registration by Photography of the Action Produced by N Rays on a Small Electric Spark”, Prosper-René Blondlot, 1904. Image credit: “Registration by Photography of the Action Produced by N Rays on a Small Electric Spark”, Prosper-René Blondlot, 1904.

Come walk with me back over 100 years to the story of N-rays, and let's find out together!

More like this

Some of the readers might enjoy reading an excerpt from Wood's letter to Nature, in which he described his visit to Blondin's lab:

I was next shown the experiment of the deviation of the rays by an aluminum prism. The aluminum lens was removed, and a screen of wet cardboard furnished with a vertical slit about 3 mm. wide put in its place. In front of the slit stood the prism, which was supposed not only to bend the sheet of rays, but to spread it out into a spectrum. The positions of the deviated rays were located by a narrow vertical line of phosphorescent paint, perhaps 0.5 mm. wide, on a piece of dry cardboard, which was moved along by means of a small driving engine. It was claimed that a movement of the screw corresponding to a motion of less than 0.1 of a millimeter was sufficient to cause the phosphorescent line to change in luminosity when it was moved across the n-ray spectrum, and this with a slit 2 or 3 mm. wide. I expressed surprise that a ray bundle 3 mm. in width could be split up into a spectrum with maxima and minima less than 0.1 of a millimeter apart, and was told that this was one of the inexplicable and astonishing properties of the rays. I was unable to see any change whatever in the brilliancy of the phosphorescent line as I moved it along, and I subsequently found that the removal of the prism (we were in a dark room) did not seem to interfere in any way with the location of the maxima and minima in the deviated (!) ray bundle.

I then suggested that an attempt be made to determine by means of the phosphorescent screen whether I had placed the prism with its refracting edge to the right or the left, but neither the experimenter nor his assistant determined the position correctly in a single case (three trials were made). This failure was attributed to fatigue.

I was next shown an experiment of a different nature. A small screen on which a number of circles had been painted with luminous paint was placed on the table in the dark room. The approach of a large steel file was supposed to alter the appearance of the spots, causing them to appear more distinct and less nebulous. I could see no change myself, though the phenomenon was described as open to no question, the change being very marked. Holding the file behind my back, I moved my arm slightly towards and away from the screen. The same changes were described by my colleague. A clock face in a dimly lighted room was believed to become much more distinct and brighter when the file was held before the eyes, owing to some peculiar effect which the rays emitted by the file exerted on the retina. I was unable to see the slightest change, though my colleague said that he could see the hands distinctly when he held the file near his eyes, while they were quite invisible when the file was removed. The room was dimly lit by a gas jet turned down low, which made blank experiments impossible. My colleague could see the change just as well when I held the file before his face, and the substitution of a piece of wood of the same size and shape as the file in no way interfered with the experiment. The substitution was of course unknown to the observer.

You can find the entire report in Nature, September 29, 1904, 530–531. There's a copy at

By Michael Richmond (not verified) on 06 Aug 2014 #permalink

Hmmm, must be all of 14 to 15 inches long. Looks impressive. Presumably the 10 to 1 reduction drive, then the plastic gears after that are for fine tuning the cavity resonator after injection of the microwaves into the TNC port at the top RHS of the pic.

Hi Ethan,

thanx for this post :) Especially nice was a link to Corey Powell's article, wish i ran into it sooner.

For all intent purposes, we can all agree the device is jack****. But I disagree with your assessment of NASA in all this.

When you say "It was not validated by NASA; it was shown by NASA to be indistinguishable from a ‘null’ device.
It does not break the laws of physics; it obeys them.", you seem to neglect that it was those same NASA people who wrote " therefore is potentially demonstrating an interaction with the quantum vacuum virtual plasma." So when they say there is no difference between 2 devices, they ARE NASA, but when they say quantum virtual plasma, they are not... :) Ghm..

I completely agree that personal views of 4 scientist working FOR NASA shouldn't necesserally be the official NASA view. But when you have an official NASA document titled "anomalous" production, and you invoke vacuum virtual plasma or whatever fairy dust, then it goes beyond personal view. The least NASA should do is check who is in charge of what, and weather credible people are running credible experiments or wasting taxpayers money on crack-pot science.

Because it was NASA's laboratory and NASA's scientist working on NASA time (not in their basement) saying that there was an anomalous thrust detected... same as OPERA saying FTL. And infact NASA is saying.."Future test plans include independent verification and validation at other test facilities.".. which is fine. But NASA isn't saying we messed up the test.

IMO, in this case NASA should be much more accountable for allowing crack-pottery to be published on it's official sites, under it's seal. Or if it's not crack-pottery to at least publish a paper explaining what is in their opinion quantum virtual plasma and how does it work.

By Sinisa Lazarek (not verified) on 06 Aug 2014 #permalink

So in NASA's defense these things are frequently based on earmarked funds put in by some congressman. Over the years I have seen several federal agencies being required (over and over) to spend money on small coherent neutrino detectors, zero pint energy devises, etc because some congressman has a constituent with an idea that will revolutionize something.

By the way there was a talk given by Irving Langmuir in the 50s on "Pathological Science" that all scientists should read and have a copy of. A PDF is available from AIP if you are a member.

I have friend whose job (in part) is to help NASA review some of the wackier submissions they get from private, well let's call them 'researchers.' The idea of bouncing EM around a cavity to produce thrust is not new. IIRC, it's probably at least 8 years old. Here's an earlier version of the idea, just for kicks and giggles.

Set up a box/cavity (a nice long one looking like a Star Trek nacelle will do). Set up two high powered lasers: fix them closer to the front bulkhead than that back, with one facing the rear and one facing the front. Turn them on simultaneously. For the millseconds between when the front-facing laser hits and the rear-facing laser hits, you have forward thrust (no, the laser set up itself is not pushing the ship back with equal force, because remember it's shooting equal lasers out in opposite directions). After the rear laser hits, force is now equal in both directions, so you are now coasting forward with a fixed velocity imparted by that few millseconds of thrust. Want to build up a high velocity? Pulse it - flick it on and off rapidly.

It works - there is nothing physically wrong with that description...except when you get to that last sentence. The problem is, the moment you turn the lasers off, the rearward facing laser will produce a deceleration equal to the acceleration produced by the foward-facing laser originally. Pulsing doesn't work; it makes the ship vibrate in place rather than move forward. So what you really have here is a very expensive and massive way to produce a very teeny tiny amount of velocity, once.

@ Eric

I can live with that :) What I am baffled by, is where in the world did they dig up or coin "quantum vacuum virtual plasma"??? :D :D

There isn't a single arXiv paper with those words used in such a way... which is a novelty in itself IMO :D

At least they could receive a credit for coining a new Sci-Fi term. Might catch-up in a new Iron Man movie... :D :D

By Sinisa Lazarek (not verified) on 07 Aug 2014 #permalink

It seems Wired answered many of your critics…

one is evident, you cannot make the difference between the null drive (which is there to confirm a theory - ie not an experiment) and a blank which is ther to challenge the experimental setup.

note also that 4 replications deserve some respect.

now, I would say simply tha the quote of Sagan is one of hs worst mistake, abused by even less cautious people.


basically theer is an insane respect for theory and a disdain for experimental result.

this result is not definitive, but fairly challenging.
no serious critic (theory is not a serious critic) is proposed that have not been considered.

only otpion is to pursue ?

if you are wrong will you say sorry ? will you resing ?
should I be right saying you are loose, as you say about those claims ?

use symmetry.

at least respect more experiment than theory, it will be better for science.

how many real innovation have been delayed of 20-60 years by theoretical arguments challenging experiments ?

A wrong experiment makes much less damage than a wrong theory that prevent continuing good experiments.
Epistemology 101.

if you feel I am a bit sharp, consider what is said in the article, and that it is a dozen time more aggressive.

By AlainCo (@alain_co) (not verified) on 07 Aug 2014 #permalink


What I am baffled by, is where in the world did they dig up or coin “quantum vacuum virtual plasma”???

Well the charitable answer is: we should ask them. If they are referring to some real, accepted phenomena and just using terms in a wierd way, asking them what they mean should clear it up. And if they prevaricate at that point, you have a stronger reason to believe they do not know what they are talking about.
Or we could just jump to the likely but more unkind answer: (they got it from) rectal extraction.

AlainCo @7:

if you are wrong will you say sorry ? will you resing ?

If the community is wrong and the drive works, I'm sure Ethan will be jumping up and down with glee like the rest of us. Because holy crap it would be nice to have such a space drive.

This is something I don't think the purveyors of these technologies really get. They see mainstream scientists as some sort of enemy, out to protect the status quo and squash their remarkable discovery. Nothing could be further from the truth. We want a space drive that doesn't need reaction mass shoved out the back. We want safe, tabletop fusion. We want cars that run on 8-volt batteries rather than gallons of hydrocarbons. The criticism you're getting is because our assessment is that you likely made an error. It is not because we don't like the idea of you succeeding.

Nothing like the good science of reading only the abstract and not the full paper. Piss-poor writing and analysis but hey way to generate click bait by proving your own article title!

This shows all the signs of pathological "science". The effect has been "replicated" but with drastically different results. The "researcher" spends a lot of time going over his CV in his video, even though that has zero bearing on validity of his claims (science isn’t authoritarian, recall that Fleischmann and Pons were “world class” chemists). And he equates the performance with the “Q” with no actual justification, and then insists that raising the “Q” from
50,000 to 1,000,000,000 will result in a thrust of 1 tonne per KW of input power. He then proceeds to launch into a space plane study showing how the cost would be reduced by more than two orders of magnitude (recall all of the benefits that Cold Fusion was going to provide). Going from “does this even work” to “look what it would mean if it worked so it must be true” is the hallmark of a con, even when self delusion is involved. The claim that it does not violate Newton’s laws and is consistent with all of the laws of physics is totally unsupported. The “Q” extrapolation very much reminds me of the not very competent physicist who measured the performance of Newman’s “energy machine” (a not particularly efficient 2 pole permanent magnet DC electric motor) and then extrapolated his incompetent measurements into an efficiency of multiple hundreds of percent (the latter being stupendously silly all by itself). The excruciating video, which is aimed at a scientifically illiterate audience, is located here:…

Point of order here.
The above article claims that a null test of the EM Drive also produced thrust.. This is simply not true.

The Null test in this case was for the Cannae drive and the difference was that the grooves on one of the plates weren't present - however both the null and active versions of the Cannae drive featured asymmetrical resonant cavities.

This strongly implies that Guido Fetta's theory is wrong, but it doesn't disprove the EM Drive.…

According to classical mechanics, reactionless drives are impossble.

Under special relativity, however, you can cheat. Relativistic linear momentum and energy are conserved in a system, but relativistic linear momentum differs from classical linear momentum, which may not be conserved.

So you can set up a closed system with microwaves in a cavity that satisfies Maxwell's equations, and conservation of relativistic energy and linear momentum, which exhibits a unidirectional force in one direction without the ejection of any reaction mass.

Papers have been written on this topic long before anyone built an actual device, and there appears to be nothing wrong with the math.

So rather than violating the laws of physics, such a machine is in perfect conformity with them.

As with all such effects, engineering them into a practical application is where the difficulties lie.

There is, however, absolutely nothing wrong with the physics.

By Eric Cordian (not verified) on 07 Aug 2014 #permalink

The EmDrive had only one version tested. It was the "virtual vacuum plasma" device - called a 'Cannae Drive' that had two devices tested. The abstract that is generally available does not make this clear, but the full paper does.

The Cannae is the one where the "functional" and "non-functional" versions were tested and both appeared to produce thrust.

The EmDrive is a different concept than that one and the paper (more of a conference notes bit than something ready for publication) reports that the EmDrive developed something like 300 micronewtons.

I do not believe either should work - the Cannae because I was taught that you can't interact with virtual particles and the EmDrive because it is a closed ('classical') system no matter how much handwaving is done about how the group velocities of the microwaves put them in a different frame of reference.

But please, get the facts straight - the team only tested one version of the EmDrive and the results of that test aren't even mentioned in the abstract that has been passed about the internet.

By Dan Hazelton (not verified) on 07 Aug 2014 #permalink

Look Ethan, I also think it extremely likely that this is a bad result, but it's just bad form to base your takedown on the abstract.

The full paper makes it clear that the null result discussed in the abstract was a test of the purported theory of the Cannae drive, not a test of the experimental setup.

They also appear to have invalidated the purported theory behind the Emdrive version, with high Q seeming to have no effect.

Having accomplished the takedown of the junk theories behind these drives, they still claim to have found an effect of some sort.

We all know there is going to be some experimental error that explains this, but if you want to publish a criticism it seems only ethical to actually read the report.

By Thomas Fisher (not verified) on 07 Aug 2014 #permalink

From Slashdot comments:

Oh for fuck's sake... Time to debunk this shit, again.
TFA got it wrong as well, so I suppose I can't blame you people for getting it wrong too, but please try doing a little more research?

A little background: The EmDrive was invented by a guy named Shawyer. It was tested by NASA, among others, and found to produce about 91 microNewtons. (I'll address the 30-50 that TFA talks about too.) That's way less than the Chinese found, but NASA was also testing it at much lower power and say they are planning to test a higher-power version.

The article mentions "... and a third person, Guido Fetta, have built three separate versions of the EmDrive". This is wrong, at least according to Fetta. Fetta invented what he calls a "Cannae Drive" ( which resembles an EmDrive but supposedly works on a different principle. In particular, Fetta believes that his drive requires radial slots in the chamber to operate. To test this, two versions of the Cannae Drive were (also, separately from the EmDrive test) tested by NASA: one with and one without the slots. Those tests both produced the same thrust (30-50 microN, about half what the EmDrive produced), which disproves Fetta's theory as to how the Cannae Drive is supposed to work.... and nothing else.

The null test device that everybody is so dismissedly claiming claiming disproves the EmDrive wasn't even supposed to be an EmDrive! Fetta, inventor of the Cannae Drive, was disproven. Shawyer, inventor of the EmDrive, was actually vindicated because according to his theory, the Cannae Drive (slots or no) is basically an inefficiently-shaped EmDrive.

I don't know why this is so hard for people to understand.

Thank you. The Fucking "Article" in the summary gets it so wrong I want to find the moron who wrote it and force him to actually read the paper that he gets almost completely wrong.

Error 1) The Cannae Drive and the EmDrive are not the same thing, at least according to the inventor of the Cannae Drive. Every result that the article talks about for the EmDrive was actually from NASA testing the Cannae Drive.

Error 2) The difference between the test and "null" devices was that one of them had slots on it (believed to be required for the Cannae Drive) and the other did not. According to Fetta (the inventor of the Cannae Drive, not just another person who built an EmDrive to test out), these slots are required. According to Shawyer (the guy who actually invented the EmDrive), they are not required. Looks like the EmDrive guy was right: they weren't required. This is addressed in Q2 of your fine link.

Error 3) TFA never mentions this, but NASA Eagleworks *ALSO* tested Shawyer's version of the drive. It was over 3 times as efficient, producing about 91 microNewtons of thrust from 17 Watts of power (the Cannae Drive got 40uN from 27W). They didn't have a "null" device for that one, aside from a resistive dummy load... which produced no thrust when energized. Also, the tested drives produced no thrust when *not* energized.

I really wish people would stop parroting the false claims in TFA.


The abstract stating that both test articles "produced thrust" is not necessarily the same as stating that both test articles "produced thrust of the same magnitude in the same direction". This article makes a big assumption in this.

It's like assuming that a CEO and a Janitor have the same wage because they both made money last year.

By Jon Minogue (not verified) on 07 Aug 2014 #permalink

Although I agree with the conclusion that there is a long long way to go before these results overcome my skepticism, I do want to warn a little against too much overreaction. This should never have hit the press; it's way too early to draw conclusions. But it's a little over the top to call it "bad" science

Science is about replication. Replication requires doing the experiment. Or, at a minimum, not laughing at other people who do the experiment.

Now: the actual results of the experiment are pretty minor. The results they show, first, didn't replicate the results that they were attempting to verify, second, falsify the actual hypothesis that they were testing, and, third, are pretty low in magnitude-- probably spurious, in my (professional*) opinion. But this is the way science is done: you test stuff. You present your results. Other scientists then critique the results, point out flaws and sources of noise and bias.

It's rather brutal, actually. But if your result holds up to the criticisms (and most don't), maybe you've pushed the boundaries of science.

These results don't-- yet. They are not yet reporting consistent results (in that their results differ significantly from those of other researchers). They have not yet eliminated possible spurious effects.

That's science.

*in fact, I am a rocket scientist

By Geoffrey A. Landis (not verified) on 08 Aug 2014 #permalink

This article, and author, are the absolute worst part of scientific endeavor. Uninformed and unread observers making hasty judgements. And to what end? To propel their own personal image of being "knowledgeable"? I'd be embarrassed if I had written this drivel. To top it off, while spouting nonsense, you are detracting from the hard-won efforts of well meaning individuals with immense knowledge of the field. And all they can do is request that their detractors attempt to read and understand what they've done, something that is I see, beyond the detractors entirely. Sadly, the public impressions espoused by these detractors sets in the minds of others, and so goes the death of promising fields. I could draw up several historical examples of such cases, in an effort to mirror your tacky use of historical context to bias your readers perspective, but I think that's generally bad form.

I won't spend my time dissecting all the fallacies of this article, but let's point to one great example. Go to the paper, and actually read what the "Null Drive" represents. The author of this article has completely misunderstood what the "Null Drive" represents. The authors of the EmDrive manuscript performed a control experiment. You may want to dig in and find it to prove to yourself the authenticity of their work. I imagine the knowledgeable peer-reviewers whom read this work with far more scrutiny than you could describe the actual experiments that were conducted. Perhaps that's why it was accepted for publication in a reputable journal?

Good luck, happy reading. I hope you actually have a subscription to the journal to access the manuscript. If you do not, I'll happily send you a copy.

By Justin English (not verified) on 08 Aug 2014 #permalink

I confess, I posted this to facebook... But I quoted this sentence in a comment that I also posted "because the effect has been replicated so many times... nobody doubts that it happens..." In hindsight, this sentence is hilarious.

*full paragraph* None of these explanations has gone unchallenged by theoreticians, and it might be fair to say that there is no accepted explanation as to how a close system of resonating microwaves can produce a thrust. There is no accepted theoretical explanation of how high-temperature superconductors work either, but because the effect has been replicated so many times, nobody doubts that it happens.

By Jason O'Hara (not verified) on 08 Aug 2014 #permalink

I have a suggestion, and I'm sure qualified people have examined it but maybe not.

People keep referring to this as a closed system.

It is NOT a closed system. It is not isolated from gravity nor is it isolated from neutrinos, to pick two examples of directional external fields against which it is currently impossible for us to shield.

So, is it plausible that something about the fields being generated within a drive chamber are interacting in a differential manner with the external fields? Call it a "Neutrino Net" if that helps. Such a differential interaction on neutrinos (for example) moving through the chamber could cause a force to be generated that averaged out to be be greater in one direction thus apparently generating thrust.

I am not a scientist but have worked with a few over the years and I read a lot of hard SF.

By Andy Dent (not verified) on 09 Aug 2014 #permalink

Sorry, but a single low res image of the outside of a tin can simply does not cut it!? I have worked with microwave guide and their associated magnetrons for nearly thirty years and without a clear view of the internal structure and waveguide form it may as well be a plant pot. Besides if microwaves are going to be used in a propulsion system surely they would serve better heating a suitable propellent which would at least be proper rocket science.

By M Paul Lloyd (not verified) on 10 Aug 2014 #permalink

I'd like to point out that theoretically, YOU CAN have thrust without ejecting any mass. Just try running a laser; you will get a (VERY SMALL) force opposite to the light beam. If you use a bank of LEDs, they might be more efficient. I don't know any theoretical reason that some microwave emitting device could not be perhaps even more efficient. Yes?

There are all sorts of ways that a force as small as 30 micronewtons could be generated via experimental error. Interaction with the Earth's magnetic field, for one. (Even a loop in the power cables driving the device could easily create the measured force.) Microwave leakage affecting the force measurement sensors for another. Yes, I'd assume they'd be aware of these possibilities and would attempt to eliminate such effects. Perhaps they were successful. But it's going to take a lot more experiments and a lot more of measurements than what we've already seen to confirm these results.

By weirdnoise (not verified) on 10 Aug 2014 #permalink

How to fool the world with bad science?

But Ethan, you aren't fooling anybody!

By John Duffield (not verified) on 11 Aug 2014 #permalink

Cheeky man, Mr. Duffield.
I'll offer 5 bucks for the scrap metal .....

I believe that you're mischaracterizing the "NULL" model. One proponent of this tech, and the designer of the Cannae drive specifically, has stated that his design works because of adaptations to the resonance chamber for his drive. Coincidentally, these adaptations make his version of the technology patentable and add value to his product.
Coincidentally, these adaptations were removed in the "NULL" apparatus.

It follows that those modifications aren't a causal factor for the propulsion behavior (and may have been a cosmetic attempt to add value to a design/attempt to jack a patent).

You inferred that the NULL device being successful indicates a problem with the experiment. I'd posit that the NULL device being successful has simply wiped away some snake oil.

By Edwina Morgan (not verified) on 12 Aug 2014 #permalink

It cannae work.

From your blog entry:

From the abstract itself:

Thrust was observed on both test articles, even though one of the test articles was designed with the expectation that it would not produce thrust. Specifically, one test article contained internal physical modifications that were designed to produce thrust, while the other did not (with the latter being referred to as the “null” test article).”…

"Figure 20. Null Test on Torsion Pendulum – average null force is 9.6 micronewtons due to 5.6A DC
current in power cable (routes power from liquid metal contacts to RF amplifier; interacts with magnetic
damper system)"

"Figure 19. Five thrust pulses of tapered RF system at 1932.6 MHz, peak thrust performance of 116
micronewtons, average thrust performance of 91.2 micronewtons"

Just an observation, Thrust was observed on both test articles, but the force was almost one magnitude higher with the real article.

By Sigurd Boaseter (not verified) on 25 Aug 2014 #permalink

Ethan, nice angel wings, love your tweets. Shawyer is a career satellite engineer. His device should be tested for the simple reason that there is so much to gain if it works.
Thankyou for bringing it to the attention of your readers and I hope you have lots of fun with explaining how it works as the experiments continue to be repeated. My ideas at viXra:1405.0252 are an attempt to explain his results from the perspective of Machian inertia, please enjoy.

By John Newell (not verified) on 03 Oct 2014 #permalink

I agree with Edwina. If you read more carefully, the "null" drive was still an experimental drive. They were testing the validity of the designers claims that his alterations were contributing to the functionality of the drive. There was still a separate control drive which did not produce any thrust.

You still make valid points about this needing to be peer reviewed and other groups validate the findings before everybody gets too excited about it, but it's also important to ensure correct understanding of the experiment before disparaging it so openly.

By Trevor Taylor (not verified) on 04 May 2015 #permalink

I'm reminded of the 'directly downwind faster than the wind' powered only by the wind *debate*:……

Accordingly, I'll hold out wishful thinking and hope that perhaps you are wrong and there is some unforseen mechanism for the effect to exist:

If I had to guess at assigning fail, I might dream up some wierd evanescent surface potential standing wave being set up along the outside of the device, producing an electric potential field thus accelerating some particular free ion cloud constituting 'surface tension' just along the waveguide (do some metal ions not *boil off* the walls of the chamber, especially in hard vacuum?).